How the seeded draw works
This is the draw to use when the result has to be shown again later. Enter the list and an integer seed, and the whole list is shuffled once; the first names become the selection and the rest stay in the table as standby order. The same seed with the same names in the same order always produces the same result, so publishing the seed together with the list lets entrants reproduce and check it themselves. Reordering even one line changes the outcome.
The shuffle is a single reverse Fisher-Yates pass, and the random numbers come from the mulberry32 algorithm, which takes a seed. The browser default random function cannot be seeded, so it cannot be reproduced. Seeds accept integers from 0 to 2,147,483,647 and the list handles up to 500 names. Because the algorithm is public, anyone can recompute the same result from the same inputs.
This is still not a cryptographic guarantee. Whoever chooses the seed in advance can also know the result in advance, so you need a separate procedure, such as publishing the seed before the draw or letting entrants set it. Goodness-of-fit testing of whether the outcomes are evenly spread is not covered here. Do not use it for an official draw that is legally binding or that involves money.
Frequently asked questions
The shuffle uses the seeded mulberry32 generator. With the same seed, the same names and the same line order, the shuffle repeats exactly. Change the seed if you want a different outcome.
No. It makes the draw reproducible and checkable, which is not the same as proving fairness. Whoever picks the seed first can know the result first, so publish the seed before the draw.
No. Goodness-of-fit tests such as chi-square are not covered. This tool only reproduces the same draw from the same seed.